| The primary motivation of our research comes from developing situational awareness (SA) technology for the applications that require rapidly assembling information and providing integrated access to and analysis of information from multiple sources. From the experiences of building these applications, we have identified a few key research areas, which form the topic of this dissertation---managing uncertain spatial information and enabling similarity search for situational awareness applications.; This dissertation has two parts. In the first part, we focus on the problem of managing uncertain spatial information for the SA applications. We propose approaches to probabilistically model and represent (potentially uncertain) event locations de scribed by human reporters in the form of free text. We analyze several types of spatial queries of interest in SA applications. We design techniques to store and index the models, to support the efficient processing of queries. Our extensive experimental evaluation over real and synthetic datasets demonstrates the effectiveness and efficiency of our approaches.; In the second part of the dissertation, we extend the spatial queries to a general class of similarity queries. A similarity search query, in general, has two components---a set of similarity predicates and a query model that combines the predicates. Based on the different application scenarios/assumptions, we propose two general similarity query retrieval and refinement frameworks to quickly retrieve relevant records that satisfy a user's retrieval need. In the first application scenario, the similarity predicates are well-defined, but the query model (combination function) may not be correct. For this case, we propose an interactive retrieval framework---I-Skyline, which combines the positive aspects of the similarity retrieval and skyline retrieval to achieve optimal retrieval in terms of the quality and the efficiency. In the second application scenario, the similarity predicates and the query model can both be wrongly specified. The primary contribution of our work is the Refinement Activation Framework (RAF) that based on a novel activation principle, it dynamically triggers the learning algorithms to quickly guide an initial query to its target form. |
